U.S. patent number 4,187,396 [Application Number 05/804,868] was granted by the patent office on 1980-02-05 for voice detector circuit.
This patent grant is currently assigned to Harris Corporation. Invention is credited to Gabriel J. Luhowy.
United States Patent |
4,187,396 |
Luhowy |
February 5, 1980 |
Voice detector circuit
Abstract
A voice detector circuit including a slew-rate limited amplifier
at the input thereof so as to reduce the interference with the
voice detection process caused by impulse noise. The slew-rate
limited amplifier is essentially transparent to audio signals,
while providing severe attenuation of impulse noise components. The
actual determination of the voice content of the audio signal is
provided by comparing the average syllabic-rate content of the
impulse noise limited, audio signal with the peak noise content
thereof. Additionally, a logarithmic amplifier is included for the
purpose of optimizing the separation of the syllabic-rate envelope
from the audio signal.
Inventors: |
Luhowy; Gabriel J. (Lima,
NY) |
Assignee: |
Harris Corporation (Cleveland,
OH)
|
Family
ID: |
25190059 |
Appl.
No.: |
05/804,868 |
Filed: |
June 9, 1977 |
Current U.S.
Class: |
704/233;
381/94.8; 455/221; 704/E11.003 |
Current CPC
Class: |
G10L
25/78 (20130101) |
Current International
Class: |
G10L
11/00 (20060101); G10L 11/02 (20060101); G10L
001/00 () |
Field of
Search: |
;179/1VC,1P,1J,1SC,1VL
;325/478 |
References Cited
[Referenced By]
U.S. Patent Documents
Other References
Electronic Design, Sep. 13, 1976, pp. 110-112, M. J. Wright, "Use
Slew-Rate Filtering". .
Radio Amateur Handbook, Apr. 1976, pp. 236, 237..
|
Primary Examiner: Gruber; Felix D.
Assistant Examiner: Kemeny; E. S.
Claims
What is claimed is:
1. Apparatus for detecting the voice content of an audio signal,
comprising slew-rate limited amplifier means responsive to said
audio signal to provide an impulse noise limited audio signal, and
having a maximum slew-rate selected such that said amplifier means
is substantially transparent to the voice content of said audio
signal, however, impulse noise components of said audio signal are
substantially attenuated thereby, and voice detector means
responsive to said impulse noise limited audio signal for
determining the voice content of said signal and providing a first
output indication thereof, wherein said voice detector means
includes syllabic-rate means for determining the average
syllabic-rate content of said impulse noise limited audio signal
and for providing a first signal in accordance therewith, threshold
means for supplying a threshold signal, and comparator means for
comparing said first signal and said threshold signal for providing
said first output indication having a value in accordance with said
comparison.
2. Apparatus as set forth in claim 1, wherein said threshold means
comprises means for detecting the peak noise content of said
impulse noise limited audio signal and for supplying said threshold
signal in accordance therewith.
3. Apparatus for detecting the voice content of an audio signal,
comprising slew-rate limited amplifier means responsive to said
audio signal to provide a noise limited audio signal, and having a
maximum slew-rate selected such that said amplifier means is
substantially transparent to the voice content of said audio
signal, however, noise components of said audio signal are
substantially attenuated thereby; and voice detector means
responsive to said noise limited audio signal for determining the
voice content of said signal and providing a first output
indication thereof, said voice detector means including
syllabic-rate means for determining the average syllabic-rate
content of said noise limited audio signal and for providing a
first signal in accordance therewith, threshold means for supplying
a threshold signal, and comparator means for comparing said first
signal and said threshold signal for providing said first output
indication having a value in accordance with said comparison,
wherein said syllabic-rate means includes logarithmic amplifier
means responsive to said noise limited audio signal for providing a
signal logarithmically related thereto for use in determining the
average syllabic-rate content of said noise limited audio
signal.
4. Apparatus as set forth in claim 1, wherein said voice detector
means further includes first timer means for delaying said first
output indication by a fixed, predetermined first delay to provide
a delayed output indication, hang timer means responsive to said
delayed output indication for providing a second output indication
which will not lapse during momentary lapses of said delayed output
indication which are shorter than a predetermined amount, and means
for deriving, from said delayed output indication and said second
output indication, a third output indication which will assume a
predetermined state after said first delay, and which thereafter
will not lapse during momentary lapses of said delayed output
indication which are shorter than said predetermined amount.
5. Apparatus as set forth in claim 1, wherein said voice detector
means further includes first timer means for delaying said first
output indication by a first amount to provide a first delayed
output indication, second timer means for delaying said first
output indication by a second amount greater than said first
amount, to provide a second delayed output indication, hang timer
means responsive to said second delayed output indication for
providing a third output indication which will not lapse during
momentary lapses of said second delayed output signal, and means
for deriving, from said first delayed output indication and said
third output indication, a fourth output indication which will
assume a predetermined state after said first delay and which,
after said second delay, will not lapse during momentary lapses of
said second delayed output indication.
6. Apparatus for providing an output signal indicating the voice
content of an audio signal, comprising syllabic-rate means for
determining the average syllabic-rate content of said audio signal
and for providing a first indication in accordance with the level
thereof, means for detecting the peak noise content of said audio
signal and for supplying a threshold signal in accordance
therewith, and comparator means for comparing said first indication
with said threshold signal to provide an output indication having a
value in accordance with said comparison, and further including
first timer means responsive to said output indication to provide a
first delayed output signal only when said output indication has
had a first value continuously for a first fixed amount of time,
said first value indicating that a voice signal is present in said
audio signal, second timer means responsive to said output
indication to provide a second delayed output signal only when said
output indication has had said first value for a second fixed
amount of time greater than said first fixed amount of time,
hangtimer means responsive to said second delayed output signal to
thereafter provide a third output signal which will not lapse
during lapses of said second delayed output signal which are
shorter than a third fixed amount, and means for deriving an output
signal from said first delayed output signal and said third output
signal, whereby said hangtimer will not be activated until after
said delay of said second fixed amount after said output
indication.
7. Apparatus for detecting the voice content of an audio signal,
comprising slew-rate limited amplifier means responsive to said
audio signal to provide an impulse noise limited signal, said
amplifier having a maximum slew-rate selected such that said
amplifier means is substantially transparent to the voice content
of said audio signal, however impulse noise components of said
audio signal are substantially attenuated thereby, syllabic-rate
detector means responsive to said impulse noise limited signal for
providing a first signal indicative of the syllabic-rate content of
said impulse noise limited signal, noise detector means for
providing a second signal indicative of the noise content of said
audio signal, and comparator means for comparing said first and
second signals to provide an output signal having a value dependent
upon the results of said comparison.
Description
BACKGROUND AND FIELD OF THE INVENTION
The present invention relates generally to a voice detector
circuit, and more particularly to a circuit which detects the
presence of a human voice in an audio signal and controls the
operation of a system (such as a squelch circuit or a channel
scanner) in accordance therewith.
Radio receivers commonly include circuits for determining when a
modulated signal is present. Circuitry of this type is often
included for the purpose of controlling a squelch gate for
selectively passing the received signal to an output only when a
modulated signal is present. Such circuits are also used in
automatic channel-scanning radio receivers for controlling the
operation of the channel scanner.
In conventional AM systems, detection of the modulated signal can
be accomplished by the simple expedient of detecting the presence
of the carrier signal. In single sideband systems, however, the
carrier is not transmitted along with the modulated information;
the presence of a modulated signal must therefore be determined by
other means. A variety of systems have been devised for this
purpose. Exemplary systems of this sort are described in the
patents to Kemper (U.S. Pat. No. 3,350,650) and Eichenberger et al.
(U.S. Pat. No. 3,102,236).
Although these circuits generally operate satisfactorily in low
noise environments, it has been found that impulse noise commonly
present in noise corrupted signals can interfere significantly with
the operation thereof.
The present invention therefore provides a system for detecting a
voice signal which includes improved noise immunity.
In accordance with the present invention, a voice detector circuit
is provided which responds to an audio signal to provide an output
indication of the presence or absence of a voice signal thereon. A
slew-rate limited amplifier is provided at the input to a voice
detector circuit for the purpose of removing impulse noise from the
audio signal. The slew-rate limited amplifier is essentially
transparent to ordinary audio signals, but greatly attenuates
impulse noise components. The voice detector circuit operates on
the impulse noise limited audio signal to provide the output
indication.
In accordance with another aspect of the present invention, the
voice detector circuit provides said output indication upon the
basis of a comparison of the average syllabic-rate content of the
audio signal with a threshold value.
BRIEF DESCRIPTION OF THE DRAWINGS
The foregoing and other objects and advantages of the present
invention will become more readily apparent from the following
description of a preferred embodiment, as taken in conjunction with
the accompanying drawings which are a part hereof, and wherein:
FIG. 1 is a schematic illustration of a voice detector circuit in
accordance with the present invention; and,
FIG. 2 is a schematic illustration of a second embodiment of a
voice detector circuit in accordance with the present
invention.
DETAILED DESCRIPTION
Referring now to the drawings, wherein the showings are for
purposes of illustrating a preferred embodiment, and not for
purposes of limiting the same, there is shown in FIG. 1 a voice
detector circuit 10 which operates on an audio signal supplied by
an audio signal source 12 to provide an output indication along an
output line 14. The audio signal source may, for example, comprise
a radio receiver detector section adapted to receive single
side-band transmissions. Obviously, audio signal source 12 may also
comprise any other source of audio signals wherein it is required
to determine the voice content thereof. The output line 14 will be
directed to the controlled system, for example a squelch gate or
channel scanner control.
The input to voice detector 10 comprises a potentiometer 16 which
is provided for the purpose of allowing adjustment of the level of
the signal supplied to the remainder of the voice detector circuit.
The wiper arm of potentiometer 16 is connected to an amplifier 18
having an output whose rate of change is limited. In other words,
amplifier 18 is a slew-rate limited amplifier. Amplifiers of this
type are described in an article entitled "Use Slew-Rate Filtering"
found on pages 110 to 112 of the Sept. 13, 1976 issue of Electronic
Design. The slew-rate of amplifier 18 is selected so that amplifier
18 is essentially transparent to voice signals, but is unable to
follow the high peak, short duration noise pulses. Therefore, a
narrow, high amplitude input square wave will be highly attenuated
by slew-rate limited amplifier 18 and will appear at the output
thereof as a much smaller triangular wave.
Slew-rate limited amplifier 18 is connected at its output to a
full-wave rectifier 20 which full-wave rectifies the slew-rate
limited signal in preparation for the detection of the syllabic
envelope. A syllabic-rate filter 22 filters the full-wave rectified
signal to provide an output signal indicative of the syllabic-rate
content thereof. The output of syllabic-rate filter 22 is, in turn,
applied to an averaging circuit 24 which averages the syllabic-rate
filtered signal over time to provide an output signal having an
amplitude which varies with the average amplitude of the
syllabic-rate content of the audio signals. The output of averaging
circuit 24 therefore indicates generally the magnitude of the voice
content of the audio signal supplied to voice detector 10.
The signal supplied by averaging circuit 24 is applied to a
comparator circuit 26 which compares the amplitude thereof with the
amplitude of a threshold signal for the purpose of providing a hard
decision regarding the presence or absence of a voice signal. In
low noise situations, this threshold signal may be derived simply
by means of a potentiometer. Generally, however, this threshold
signal will be developed by means of a high-pass filter 30 and peak
detector 32 so as to provide a threshold signal whose amplitude
indicates the peak noise content of the audio signal. This approach
is preferable in high noise environments because the noise content
of the input signal will decrease when a modulated signal is
supplied by the audio signal source 12, due to the action of the
automatic gain control loop which will generally be included in
audio source 12. Consequently, not only will the output of
averaging circuit 24 increase, but also the output of peak detector
32 will decrease, thereby providing more rapid and positive
switching by comparator 26. In any event, the output of comparator
26 will indicate whether or not the syllabic-rate content of the
audio input to the voice detector circuit exceeds a predetermined
amount.
The comparator output signal is supplied to a timing circuit
generally indicated by reference number 34, whose purpose is to
provide an output indication which does not lapse during ordinary
pauses in the voice signal.
Timing circuit 34 includes a 30 millisecond timer circuit 36 which
delays the actuation of the output indication provided along output
line 14 until 30 milliseconds after the output of comparator 26 has
begun to indicate that a voice signal is present. This delay is
included to further reduce the susceptability of the voice detector
circuit to impulse noise falsing. If the output of comparator 26
continuously indicates that a voice signal is present for at least
30 milliseconds, then the probability is quite high that a true
voice signal has been detected. The 30 millisecond timer will
therefore provide a logic "1" output indication (indicating that a
voice signal has been detected) when 30 milliseconds have elapsed
since the output of comparator 26 last indicated that no voice
signal was present. The signal will be gated by OR gate 40 to the
output line 14.
A hang-time timer 42 is provided for the purpose of preventing the
indication provided along output line 14 from changing during short
pauses in the voice signal. Normally, these pauses would cause the
output indication to momentarily indicate that no voice signal was
present. Hang-time timer 42, however, will continue to provide a
logic "1" signal to OR gate 40 for a short period of time
(approximately 1.2 seconds) following the conclusion of each voice
segment. Although the hang-time timer 42 could be actuated by the
output of 30 millisecond timer 38, it will preferably be actuated
by a 50 millisecond timer 44 instead. This longer period of time
(50 ms) is selected so as to more certainly assure that a voice
signal has indeed been detected.
The 30 and 50 millisecond times associated respectively with timers
38 and 44 have been selected since it is a characteristic of speech
waveforms that if 30 milliseconds or less of the first speech
utterance is lost, the loss is almost unnoticeable. As stated
previously, however, the delay of 30 milliseconds does drastically
reduce the problem of impulse noise falsing. The 50 millisecond
time length is selected because the probability that speech is
truly present approaches 98% if an utterance lasts longer than 50
milliseconds.
A second embodiment of the invention is shown in FIG. 2. For
simplicity of description, elements of FIG. 2 which correspond to
identical elements in FIG. 1 are denoted by similar reference
numerals. This embodiment is quite similar to that shown in FIG. 1
except that a log amplifier 46 is interposed between the full-wave
rectifier 20 on the one hand and low-pass and high-pass filters 22
and 30 on the other hand. This log amplifier is included for the
purpose of simplifying the separation of the syllabic-rate envelope
from the voice "carrier." This may be understood more readily when
it is appreciated that the audio signal supplied to slew-rate
limited amplifier 18 is essentially an amplitude modulated signal
which may be characterized as a multiplicative combination of a
carrier signal and a modulating signal. By including a log amp in
the manner shown in FIG. 2, this multiplicative combination is
changed instead to an additive combination of the carrier signal
and the modulating signal. Consequently, a better signal-to-noise
ratio is theoretically possible, thereby making it easier to
separate the syllabic-rate envelope from the voice carrier.
FIG. 2 also illustrates a timing circuit 34' wherein the hang-time
timer 42' is operated directly from the 30 millisecond timer 38',
as discussed previously.
Although the invention has been described with respect to a
preferred embodiment, it will be appreciated that any number of
modifications may be made therein without departing from the spirit
and scope of the invention as defined in the appended claims.
* * * * *